This demo is intended to illustrate the kinds of applied analysis reports that will be well within reach after the BootCamp.

Load packages

These are the extra tools (beyond the default R functionality) we’ll be using in this report.

library(plotly)
library(DT)
library(data.table)
library(tidyverse)

Load data

Let’s load the following files from the DepMap 21Q2 data release, all of which can be downloaded from the depmap portal:

For the gene expression and dependency data we’ll also extract data for just a single gene (MDM2).

#cell line metadata
cell_line_meta <- read_csv('data/sample_info.csv') 

#load gene expression data and keep only the cell line IDs and the MDM2 expression levels
CCLE_GE <- fread('data/CCLE_expression.csv') %>% 
  as_tibble() %>% 
  select(DepMap_ID = V1, MDM2_GE = `MDM2 (4193)`)

#load the Achilles data and keep only the cell line IDs and the MDM2 dependency scores
Achilles_dep <- fread('data/Achilles_gene_effect.csv') %>% 
  as_tibble() %>% 
  select(DepMap_ID = V1, MDM2_dep = `MDM2 (4193)`)

Make table

Use only the cell lines labeled as metastatic samples, and combine the metadata, selected expression and dependency data all into a single table

cell_line_data <- cell_line_meta %>% 
  filter(primary_or_metastasis == 'Metastasis') %>% 
  select(DepMap_ID, CCLE_Name, lineage, lineage_subtype) %>% 
  inner_join(CCLE_GE, by = 'DepMap_ID') %>% 
  inner_join(Achilles_dep, by = 'DepMap_ID')

Make a table showing the average MDM2 dependency and mRNA expression for each lineage

cell_line_data %>% 
  group_by(lineage) %>% 
  summarise(avg_MDM2_dep = mean(MDM2_dep, na.rm=TRUE),
            avg_MDM2_GE = mean(MDM2_GE, na.rm=TRUE)) %>% 
  arrange(avg_MDM2_dep) %>% 
  datatable(rownames = FALSE)

Make plot

Make an interactive scatterplot of MDM2 expression and dependency status across these cell lines

g <- ggplot(cell_line_data, mapping = aes(x = MDM2_GE, y = MDM2_dep, color = lineage, text = CCLE_Name)) + 
  geom_point() +
  labs('MDM2 GE (log2(TPM))', y = 'MDM2 dependency')
ggplotly(g)